Ada Lovelace has much to teach us about AI

On Ada Lovelace Day, Battle of Ideas festival producer Sandy Starr discusses the life and work of this self-effacing pioneer, and argues that she can help us get to grips with today's AI.

Today is Ada Lovelace Day, when people remember the nineteenth-century computing pioneer Ada Lovelace and also celebrate women's contributions to science and technology. What can Lovelace's life and work teach us today?

She was born Ada Byron in 1815, daughter of the colourful public figures Lord and Lady Byron (a great Romantic poet and a leading educational reformer respectively, among their many other claims to fame). Ada Byron became Ada King when she married the politician and scientist William King in 1835, then she became Countess of Lovelace when William was made Earl of Lovelace in 1838.

Ada Lovelace was a gifted mathematician, at a time when mathematics was widely assumed to be the exclusive province of men. One might conclude from this that she had to fight to defy stereotypes. The truth is actually more complex than that.

Lady Byron, herself a lover of mathematics, tried to use mathematical education to dissuade Lovelace from taking after her errant poet father. Lovelace duly came to love mathematics in turn, but she pushed back against her mother's overcorrection, seeking to pursue a more integrated approach that she referred to (in a letter to her mother) as 'poetical science'.

Lovelace collaborated with the English polymath Charles Babbage on his Analytical Engine, a precursor to the modern electronic computer (it was never actually built). A cursory glance at Lovelace's work on the Engine might conclude that she made a great contribution to the project, only to be sidelined by prevailing sexism. Again, the truth is more complex.

Lovelace's legacy has had to be rescued not just from the sexism of her time (although that is a factor), but also from her own self-effacement, which could be excessive even by nineteenth-century standards. This self-effacement served as an important counter to Babbage's foibles, which were an impediment to the expression of his ideas.

Nobody would describe Babbage as self-effacing, but he was too fussy and precious to present a thoroughgoing explanation of the Analytical Engine in public. Lovelace therefore stepped in and did it for him. The way she was forced to go about this was ridiculously convoluted, involving greater patience than might be expected even of a dedicated amanuensis.

Lovelace met and corresponded with Babbage in England, but instead of collaborating with him directly on a paper about the Analytical Engine, she ultimately had to wait until he went to Turin and gave his only ever detailed talks about the device. These talks were then summarised in French by an Italian engineer (Luigi Menabrea, who would go on to become prime minister of Italy) and the resulting published summary was then translated back into English by Lovelace.

To this translation, Lovelace appended copious Notes by the Translator. These 'Notes' included a large foldout page, containing a table that has some claim to being the world's first ever computer program (it isn't quite that, but it's not far off). It's in these 'Notes' that Lovelace fleshed out Babbage's ideas and helped him to establish a basis for general-purpose computing.

The 'Notes' as originally published were attributed solely to 'AAL' (the initials stand for 'Ada Augusta Lovelace'), and Lovelace reportedly had to be persuaded to take even this cryptic credit for having written them. Seeking recognition for herself seems not to have been at the forefront of her mind. Rather, she was trying to perfect and publicise a brilliant invention, rescuing it from the reticence of its inventor.

None of this is to begrudge the recognition that has been given belatedly to Lovelace's work. If anything, some of the ideas expressed in her 'Notes by the Translator' are more relevant 180 years later than even her admirers appreciate. In particular – as I argue in my Letter on Liberty, AI: Separating Man from Machine – these ideas can help us to think through the conundrums of today's artificial-intelligence technologies.

In his landmark paper exploring the question 'Can machines think?', Alan Turing – the pioneer of modern electronic computing – engaged with what he famously called 'Lady Lovelace's Objection'. This refers to Lovelace's caution, in her 'Notes', that the Analytical Engine is not a creator of entirely original work – that it 'has no pretensions whatever to originate anything', and that 'its province is to assist us in making available what we are already acquainted with'.

But Lovelace's view of the matter was more nuanced than these statements might suggest. She argues elsewhere in the 'Notes' that if certain preconditions can be met, the Analytical Engine 'might compose elaborate and scientific pieces of music of any degree of complexity or extent'. This was an astonishing thing to suggest in 1843, going beyond the self-playing instruments that existed at the time and imagining musical composition via computer.

Even setting aside Lovelace's explicitly artistic speculations, one gets the impression that she thought of the mathematical uses of the device – the fact that, as she puts it, 'the Analytical Engine weaves algebraical patterns' – as a form of artistry, or 'poetical science', in itself. This sort of outlook anticipates the title of computer scientist Donald Knuth's magnum opus, The Art of Computer Programming.

Lovelace occupied, and tentatively began to map, a strange and rich space of possibilities that sits between art and numbers. It is from such a space that twenty-first-century 'generative' AI like ChatGPT – which can generate not just everyday language, but also programming code, when prompted – eventually blossomed. It is for good reason that two of the large language models in the GPT-3 family (now being supplanted by GPT-4) were named 'Ada' and 'Babbage'.

Lovelace seemed content to be characterised as midwife to Babbage's ideas. Such dedication to recognising and realising great ideas is laudable, but she was overly modest. It is therefore right that we give Lovelace the credit she neglected to give herself, although we should be careful not to oversimplify her story in the process.

Sandy Starr is deputy director of the Progress Educational Trust. He is chairing the debates Terminator or tech hype? AI and the Apocalypse, What are the limits of AI? and Is AI the end of art? at the Battle of Ideas festival in London on 28 & 29 October. He is also chairing the debate What are the limits of AI? at Buxton Battle of Ideas festival on 25 November.

To find out more about the Battle of Ideas festival and to buy tickets, visit the festival website. Free subscribers get 10% off tickets with promo code SUBSTACK-BOIF23. Paid subscribers can get even bigger discounts by using our Associate rates.